Corrosion Attack Research Articles

Life assessment of thermal sprayed alloy 825 in waste heat incineration environment

In this study, the cyclic hot corrosion behaviour of Ni-based superalloy 825 before and after thermal spray coating of Cr3C2–25%NiCr and NiCrAlY powders was investigated in a superheater section of a simulate waste-heat incinerator. The thermogravimetric (TG) technique is used to analyze the corrosion kinetics of the samples that have undergone hot corrosion tests in Na2SO4–40%K2SO4–10%KCl–10%NaCl environment at 750 °C for 60 cycles. The X-ray diffraction (XRD) technique was also used to analyze the phases of coating and corrosion products. Scanning electron microscopy (SEM) was used to characterize the surface and cross-sectional morphology of the coatings, while the compositions were analyzed using energy dispersive spectroscopy (EDS). After hot corrosion tests, the cross-sections of the coated samples were investigated using EDS point and line analysis techniques to characterize the oxide scales formed and assess the depth of the corrosion attack. It is confirmed that the bare specimen showed maximum weight gain compared to the HVOF coated specimens. Minor spallation has been observed in both the uncoated and specimens coated with the Cr3C2–25%NiCr powder, however, not in the NiCrAlY coated specimen. Based on the weight gain results, the Cr3C2–25%NiCr coating exhibits less weight gain than the NiCrAlY coating. However, some failures were noticed in the leading edges of Cr3C2–25%NiCr coated samples with minor spallation and sputtering. The NiCrAlY powder-coated specimen displayed better corrosion resistance than the specimens coated with Cr3C2–25%NiCr powder due to the development of NiO, Cr2O3, Al2O3, and NiCr2O4 scales on the surface of hot corroded samples. The NiCrAlY coating was found to have 20% reduced weight gain compared to the uncoated alloy.

THE STUDY OF THE CAUSE OF FAILURE OF THE MAIN ELEMENTS OF THE CHIRTMAS TREE AND EFFECTIVE SOLUTION WAYS

The article is devoted to the study of the causes of failures of fountain fittings in the process of operation, the measures taken to eliminate failures, increase efficiency and increase the overhaul period. İn order to improve the efficiency of the fountain fittings, the works that can be performed during major and current repairs are investigated. The main purpose of this article is to increase the reliability of the design of the fountain fittings. This article explores the causes of failure that may occur in wellhead christmas trees, their causes and solutions. The general design of x-mas trees is shown. Basically, christmas tree fittings are cross-shaped, three-necked and, depending on the design, for a working pressure of 21, 35, 70, 105, 140 MPa. x-mas trees perform the main and most important work, such as hanging tubing, shutting off the wellhead, providing the necessary pressure on the wellhead, and regulating oil production. As is known, during the operation of oil and gas wells, along with the main product, clay, sand and abrasive particles are removed from the well. Among these impacts, christmas trees are exposed to the corrosive attack of sand flowing with oil, gas and water during operation. It is the cross and the connecting elements of the x-mas tree - closing valves, the cross, the christmas tree, a pair of gates - saddles, the sealing elements that are most susceptible to eating (wear). Due to the fact that these processes occur under high pressure, a number of important requirements are imposed on the design of x-mas trees. İmproving the design of x-mas trees, which is wellhead equipment, is always one of the most important tasks in operation. In order to improve the efficiency of x-mas trees, the works that can be performed during major and current repairs have been studied. The main purpose of this article is to study the improvement of the reliability indicators of the x-mas tree design, increase the overhaul period. Keywords: Christmas tree, stem, valve, globe valve,packing element, shut-off valve, disc - saddle pair.

A Systematic Study on the Effects of Key Influencing Factors on a Pull-Off Adhesion Test of Organic Coatings

Coatings can protect metallic substrates (e.g., piping, vessels, tubes, etc.) against corrosion attack. The adhesion strength of a coating to its substrate is an essential factor that defines the overall coating integrity and performance. A pull-off adhesion test is commonly performed to determine the coating adhesion strength. This study evaluated the effect of various vital influencing factors on the pull-off adhesion test, including surface preparation of the loading fixture (dolly) and coating surface, types, curing durations and conditions of adhesive, and scoring of coating around the dolly. Test results were systematically analyzed with nonsubjective tools such as ImageJ software for failure mode assessment and statistical analysis (t-test) to compare adhesion values.

Achieving high tribological and corrosion performances via a new approach: Al2O3/CrN duplex coatings

Owing to growth defects, the CrN coating–substrate system may suffer from severe corrosion and wear attack in a marine environment, which limits the coating protection against the 316L substrate. To improve the efficiency, a new approach of fabricating Al2O3/CrN duplex coatings to achieve excellent corrosion and wear performances was provided by sealing the growth defects and extending the deposition time of the CrN layer. The Al2O3/CrN duplex coatings were obtained by multi-arc ion plating system sputter CrN layers combined with sequential atomic layer deposition Al2O3 layers. The morphology and structure of the as-deposited coating were investigated by scanning electron microscopy and x-ray photoelectron spectroscopy. The corrosion performances and tribological properties of the Al2O3/CrN duplex coatings in a seawater environment were investigated. The results showed that the density of the duplex coating decreases as the CrN layer thickness increases. The Al2O3/CrN duplex coatings with a suitable CrN layer thickness exhibited superior corrosion resistance in a seawater environment, which may be ascribed to the Al2O3 layer filling or sealing the partial growth defects of the CrN layer and playing a good barrier role to prevent the infiltration and penetration of corrosive media. Furthermore, the wear rate of the Al2O3/CrN duplex coating was reduced more than 30% compared to the single CrN coating of the same thickness. The good wear resistance of Al2O3/CrN duplex coatings was attributed to the synergistic effect of the dense structure, high wear resistance of the Al2O3 layer, and the supporting effect of the CrN layer as well as the formation of the lubricating component of CaCO3 and Mg(OH)2 during wear progress.

Investigation of adsorption, corrosion inhibition, synergistic inhibition effect and stability studies of Rheum ribes leaf extract on mild steel in 1 M HCl solution

BackgroundDue to high cost, toxicity of conventional inhibitors and environmental regulations, corrosion scientists have directed toward plant extracts as alternatives to conventional inhibitors. In this study, leaf extracts of Rheum ribes (RR leaf) were prepared, and their adsorption and protective effects on mild steel (MS) corrosion were investigated in 1.0 M HCl solution. The synergistic inhibition effect of iodide ions and the stability of surface inhibitor film were also studied. MethodsThe protection ability of extract was investigated using electrochemical techniques. The stability of surface film was evaluated by potentiodynamic and potentiostatic studies. After being exposed to the corrosive medium, the surface of metal was examined using SEM, EDX, AFM and contact angle measurements. Significant FindingsThe extract molecules formed an adherent and homogenous protective film on the MS surface. When the RR leaf extract concentration increased, a stronger inhibition effect against corrosive attack was obtained. At 1000 ppm extract concentration, the protection efficiency of RR leaf extract increased from 91.8% to 94.9% when 1000 ppm KI was added to the inhibited acid solution. The extract molecules adsorbed to the metal surface via physical and chemical interactions. The surface extract film was very stable under potentiostatic or potentiodynamic conditions.

Seismic Retrofit of Nonuniformly Corroded Coastal Bridge Piers with FRP and Engineered Cementitious Composite Overlays

Coastal bridge piers are prone to corrosion attack owing to their exposure to marine environments, particularly in the splash and tidal zone, which is rich in chloride, oxygen, and moisture. Severe corrosion in this area results in the transfer of plastic hinges and evidently impairs the structural performance under seismic loading. Previous seismic retrofitting typically focused on the area of the pier end with the greatest bending moment. However, the location of the splash and tidal zone varies with the change in season and is generally not in the pier end. This study investigated for the first time the seismic retrofit of such piers using fiber-reinforced polymer (FRP) and engineered cementitious composite (ECC), where the middle part of the pier subjected to NaCl solution corrosion was simulated as the splash and tidal zone. Reversed cyclic loading tests of nine bridge piers with various corrosion rates, FRP types, and retrofit areas were performed. The seismic behaviors in terms of failure mode, crack pattern, load-carrying capacity, ductility, energy dissipation capacity, and stiffness degradation were compared and discussed. The experimental results of mildly corroded specimens indicated that seismic retrofitting with combined FRP and ECC significantly improved the seismic performance of bridge piers. The retrofitted specimen exhibited an even higher ductility and energy dissipation capacity than the noncorroded specimen. For severely corroded piers, the plastic hinge was transferred from the pier end to the corroded area and then transferred back to the pier end after retrofitting. The seismic performance of the specimens retrofitted from the corroded area to the foundation was better than that of the piers retrofitted only in the corroded area.

Corrosion behaviour of MgO-based refractories by different existence states of manganese-containing volatile phases

Medium/high manganese steel is great interest because of its excellent energy absorption and strength. Further, it has emerged as a potential advanced high-strength steel in automobiles. In this study, the corrosion behaviour of MgO-based refractories by different existence states of manganese-containing volatile phases was systematically investigated via laboratory experiments of the vacuum treatment of medium/high manganese steel. The dissection of a magnesia crucible indicated that the manganese-containing volatile phases exist in three states: manganese vapour, liquid manganese and manganese oxides. The corrosion mechanism models of the MgO-based refractories were established as follows: (1) For the manganese vapour, the periclase was dissociated into the fine blocks when the magnesia-carbon brick was continuously exposed to the manganese vapour. (2) For the liquid manganese, a MnO layer was detected near the manganese metal, which was resulted from the MgO oxidation, and a further reaction took place to form a MnO·MgO solid solution layer at elevated temperatures, which led to the poor binding ability of the MgO crucibles. (3) For the manganese oxides, the Al2O3 crucibles were preferentially corroded by the molten manganese oxides, which are attributed to the superior thermodynamic driving force for the formation of MnAlO4. Moreover, a reaction layer composed of a MnO·MgO solid solution and a low-melting-point silicic acid phase was formed at the interface between the manganese oxides and the magnesia-carbon brick along the decarburized infiltrated zone. In addition, the molten manganese oxides infiltrated along the periclase grain boundaries into the magnesia-carbon brick and to form a reactive metamorphic layer. Consequently, under the attack of physical and chemical corrosion, the refractory particles were loosened and dropped into the corrosive manganese oxides, causing the irreversible damage of the refractory. This work provides a strong theoretical basis for the qualitative evaluating damage of MgO-based refractories by manganese-containing volatile phases.

Corrosion Effects on the Mechanical Properties of Spun Pile Materials

Prestressed concrete piles with closed-ended circular hollow sections (spun piles) are sometimes used as foundations for pile-supported wharves. Due to a reduction in the rebar area, concrete compressive strength, yield strength of PC-bar, and bond strength between PC-bar and concrete, corrosion attacks typically lower the performance of spun piles in the marine environment. A comprehensive analysis of the corrosion effect on the mechanical properties of the spun pile materials is crucial to assess the performance of corroded spun plies. Using a three-dimensional finite element analysis (FEA), this study aimed to evaluate the impact of corrosion on the mechanical properties of the material used in spun pile construction. We simulated the effect of nonuniformly distributed corrosion products using a volumetric strain expansion over 0–75 years. The FEA results provided the stress–strain relationship of the corroded spun pile materials and the bond–slip relationship between the corroded PC-bar and concrete. We proposed equations for predicting the deterioration degree of the mechanical properties of corroded spun pile materials and compared them to those presented in previous studies. It was shown that the compressive strength of the corroded cover concrete decreased significantly after the corrosion degree reached 12%, which the previous research had not expected. The bond strength reduction was inverse exponential against the elapsed years after the corrosion degree reached 1.3%. Moreover, the yield strength of PC-bars decreased linearly with the increase in the corrosion degree.

Mechanistic Model with Empirical Pitting Onset Approach for Detailed and Efficient Virtual Analysis of Atmospheric Bimetallic Corrosion.

A mechanistic model of atmospheric bimetallic corrosion with a simplified empirical approach to the onset of localized corrosion attacks is presented. The model was built for a typical bimetallic sample containing aluminum alloy 1050 and stainless steel 316L sheets. A strategy was developed that allowed the model to be calibrated against the measured galvanic current, geometrical corrosion attack properties, and corrosion products. The pitting-onset simplification sets all pits to be formed at a position near the nobler metal and treated all pits as being of the same shape and size. The position was based on the location of the highest pitting events and the pit attributes on an average of the deepest pits. For 5 h exposure at controlled RH (85%, 91%, and 97%) and salt load (86 μg NaCl/cm2), the model was shown to be promising: both for analysis of local bimetallic corrosion chemistry, such as pH and corrosion products, and for efficient assessment of pitting damage by computing a single largest pit depth. Parametric studies indicated that the pitting-onset approximation deviated the most at the beginning of exposure and when RH was below 91%.

Influence of Inclusions in the Corrosion Behavior of Plasma Nitrided Stainless Steel

Plasma nitriding is a well‐established technique to improve hardness and tribological properties of austenitic stainless steel. It is proved that it is also possible to preserve the corrosion resistance after nitriding by controlling the process parameters such as time and temperature, obtaining the so‐called S phase. Herein, the corrosion behavior is evaluated for nitrided layers produced by three different plasma treatments: DC plasma nitriding, plasma immersion ion implantation (PIII), and low‐energy ion implantation (LEII), using different parameters in order to obtain the S‐phase in thickness from 1.2 to about 6 μm without nitride precipitation. The microstructure and chemical composition of the nitrided layers is characterized by means of X‐ray diffraction, secondary‐ion mass spectroscopy, and scanning electron microscopy–focused ion beam. The corrosion behavior is evaluated by means of the cyclic polarization tests in NaCl solution. The morphology of the corrosion attack is studied by optical microscopy and SEM‐FIB, revealing a change from crevice to pitting after the nitriding process. The inclusions are observed to be corrosion initiation sites. Due to this, the thickest nitrided layers (with high N concentration) show better corrosion behavior than the thinner ones.

Laser Obtained Superhydrophobic State for Stainless Steel Corrosion Protection, a Review

Stainless steel has become an integral part of modern engineering materials and daily life due to its mechanical efficiency, strength, recyclability, high resistance to oxidation and corrosive attack, which make it the ideal material for many kinds of applications. At the same time, steel suffers from certain types of corrosion, such as intergranular corrosion, or contact corrosion that develops when stainless steel comes into contact with carbon steel or another metal with a different electrochemical potential. Finally, pitting corrosion is a serious problem often occurring when stainless steel parts work in sea water. This paper provides a brief overview of methods for protecting stainless steel from corrosion using a new approach based on superhydrophobization of the surface of stainless steel using laser processing followed by the deposition of a layer of a substance with a low surface energy. The review discusses the mechanisms of corrosion protection by such coatings and the properties of superhydrophobic coatings presented in the literature. Superhydrophobic protective coatings on stainless steel have been shown to significantly reduce corrosion, with some demonstrating a decrease in corrosion current of up to 156 times. However, a more comprehensive analysis of the mechanisms contributing to this effect, as well as a comparison with anti-corrosion coatings on other metals, suggests that the combination of these mechanisms has the potential to create even more durable and effective surfaces for corrosion protection of stainless steel.

Effects of cracks on chloride‐induced corrosion initiation and propagation of carbon and stainless steel rebar

AbstractCurrent durability design models for reinforced concrete structures in chloride‐contaminated environments do not consider the effects of cracks, which in practice are almost inevitable. Literature reports controversial results on corrosion propagation time of carbon steel (CS) in cracked concrete, while fewer information can be found on stainless steel (SS) rebar, which can be employed as additional protection strategy. In this experimental study, corrosion initiation and propagation of CS and SS rebar in uncracked and cracked concrete are presented. Prismatic specimens, reinforced with CS and 304L SS bars were subjected to a loading procedure to induce longitudinal micro‐cracks and exposed to a 3.5% NaCl solution for more than 2 years. Corrosion was monitored with electrochemical techniques and at the end of exposure corrosive attacks were observed. Results demonstrate the fundamental contribution of micro‐cracks in accelerating corrosion phenomena for CS, while SS is still in passive conditions also in cracked concrete.

Experimental Study on Mechanical Properties and Durability of Polymer Silica Fume Concrete with Vinyl Ester Resin.

Polymer concrete, which contains silica fume powder and vinyl ester resin as two replacements for Portland cement, has improved mechanical properties and durability compared to ordinary concrete. Thus, this kind of concrete is considered to be a high-strength concrete that is resistant to corrosion and chemical attacks. In this paper, the effects of the combination of silica fume powder and vinyl ester resin as two Portland cement replacements on the workability and slump value, initial and final water absorption, compressive and tensile strength, and failure and fracture paths of the polymer concrete have been investigated. All investigations have been based on 16 different polymer concrete mixture designs. The results indicate that the optimum percentages for a combination of silica fume and vinyl ester resin, which has the maximum compressive strength (34.26 MPa) and the maximum tensile strength (4.92 MPa), are a combination of 10% silica fume and 5% vinyl ester resin. To evaluate the durability of polymer concrete, the water absorption of all mixture designs has also been measured. Accordingly, the mixture design, which includes a combination of 15% vinyl ester resin and 5% silica fume, has a minimum initial and final water absorption equal to 0.62% and 1.95%, respectively.

Adsorption and corrosion inhibition capability of Rheum ribes root extract (Işgın) for mild steel protection in acidic medium: A comprehensive electrochemical, surface characterization, synergistic inhibition effect, and stability study

The adsorption and inhibitory effect of Rheum ribes (RR) root extracts on mild steel (MS) corrosion in 1 M HCI solution were examined by electrochemical techniques. The characteristics of the electrode surface after exposure to corrosive solutions were investigated using scanning electron microscopy (SEM) and atomic force microscopy (AFM). The chemical composition and the distribution of the surface film was examined with energy dispersive X-ray spectroscopy (EDX). The hydrophilic or hydrophobic properties of the MS surface were investigated using contact angle (CA) measurements. According to the findings, molecules of RR root extract adsorb to the metal surface, which form a protective and uniform film. Increasing the concentration of RR root extract improves the resistance of the alloy's surface to corrosion attacks. The extract acts as a mixed-type corrosion inhibitor with dominantly cathodic action. The molecules of the extract predominate the anodic reaction in the presence of iodide ions. Physical interactions are the most important in the extract molecule adsorption. The synergy between the iodide ions and the molecules of the extract improves the extracts protection ability and the film's stability. The film of extract molecules on the surface of the steel reduces surface deterioration and increases hydrophobicity.

Material compatibility of SI engine components towards corrosive effects on methanol-gasoline blends for flex fuel applications

The experimental study is conducted through static immersion of engine components to evaluate the material compatibility against the corrosive nature of methanol. The investigation is performed at room temperature along with different fractions of methanol in gasoline (M0, M20, M50, and M100) for varied durations of 60, 120, and 180 days. Further, the surface characterization of metal samples and fuel properties are examined in previous and post-immersion test duration in order to comprehend the penetration of corrosion. Also, the impact of corrosion on morphological changes, composition and engender of oxide layers on the engine components are inspected by SEM/EDX and XRD. Similarly, the changes brought by corrosion products on the physical (density and viscosity) and chemical properties (calorific value and TAN) are investigated and compared before immersion. The result of the study demonstrates that increasing the methanol percentage intensifies the corrosiveness of fuel samples when exposed to metal samples. Additionally, it has been found that the presence of oxygen in methanol significantly contributes to an increase in metal corrosion. Based on the investigational outcomes, comparatively the piston exhibits lesser resistance towards corrosion than the piston ring and valve. On the other hand, corrosion attack on the engine component surfaces disclosed to M100 is higher relative to that of M50, M20 and M0. Consequently, the characteristics of fuels are changed with respect to corrosion intensity and M100 has a higher variation among fuel samples. Therefore, the application of methanol in a lower concentration is suitable for SI engines in accordance with material compatibility since the corrosiveness of methanol is minimal on the engine components. Finally, this work reports some proposals based on the experimental outcomes to exalt the performance of engine components against the corrosive nature of methanol for flex-fuel engines (FFE).

In Vitro Investigation of Corrosion Control of Magnesium with Degradable Polycaprolactone Coatings for Cardiovascular Grafts

Magnesium is a promising metal for resorbable cardiovascular implants due to its high biocompatibility, high corrosion tendency, and mechanical properties. However, adapting its corrosion rate to the physiological healing processes is required to ascertain a safe graft function. A protective polymeric layer is supposed to slow down the corrosion rate of magnesium. Additionally, coatings can improve the host’s tissue interaction with the implant by implementing the local delivery of antibiotic drugs and growth or cell adhesion factors. However, little is known about the interaction of polymer-based coatings, their degradation, and magnesium corrosion. This study examines the corrosion mechanism of magnesium protected by spin coatings and electrospun fiber coatings under physiological conditions. Pure magnesium specimens were coated with polycaprolactone (PCL). The corrosion of the coated magnesium was evaluated using an immersion test in simulated body fluid. Spin coatings provided efficient protection against corrosive attacks and a significantly lower corrosion rate by 75% compared to uncoated magnesium. In contrast, fiber coatings did not provide relevant corrosion protection. On the other hand, magnesium corrosion caused the accelerated degradation of the PCL layer. A reliable and safe implant function is vital, especially in cardiovascular applications. Magnesium coating, therefore, should be carried out with spin coatings.

A material degradation study of novel FeCrAl alloys, stainless steels and nickel base alloy in fluidized bed heat exchangers of a waste-fired CFB boiler

This study aimed to evaluate the material degradation resistance of two newly developed FeCrAl alloys exposed within the loop seal region of an 85 MWth waste-fired CFB boiler in Sweden. In addition to the FeCrAl alloys, two commercial austenitic stainless steel and one nickel base (Ni-base) alloy were also studied. The samples were exposed for 6 and 12 months by clamping half-moon rings onto tubes of the installed superheater bundle, achieving a material temperature of 500–520 °C. The material degradation of the samples was assessed by evaluating material loss using an ultrasonic thickness gauge in combination with a scanning electron microscope (SEM). Microstructural analysis was carried out using SEM coupled with Energy-dispersive X-ray spectroscopy (EDX).Both FeCrAl alloys showed promising results achieving material losses in the same range as the investigated austenitic stainless steels and the Ni-base alloy. A thin inward-growing Cr/Al-rich and thick outward-growing Fe-rich oxide were present for the FeCrAl alloys, and an internal nitridation zone was formed in the material matrix close to the metal/oxide interphase. The material loss and extent of the corrosion attack for the austenitic stainless steels varied depending on the alloy composition. However, the corrosion attack remained similar, as all the austenitic steels suffered from internal and intergranular corrosion. For the Ni-base alloy, the attack was defined by an internal Cr-rich oxide formation with no intergranular corrosion.This study suggests that the novel FeCrAl alloys provide satisfying resistance towards corrosion and erosion of the fluidized bed heat exchangers in the loop seal region of a waste-fired boiler. In addition, the stainless steel SX and the Ni-base alloy Sanicro® 69 performed well. Furthermore, the results revealed that corrosion rather than erosion was the dominating degradation process of the investigated samples, as deposit formation was observed on most of the samples as well as corrosion product layers. Thus, the erosive aspect of the attack was expected to be minor.

Evolution of degradation mechanism and fixation strength of biodegradable Zn–Cu wire as sternum closure suture: An in vitro study

This work reports the first in vitro study on the in-situ biodegradation behaviour and the evolution of fixation strength of Zn–Cu alloy wires in a simulated sternum closure environment. Zn–Cu wires were used to reapproximate the partial bisected sternum models, and their fixation effect was compared with traditional surgical grade 316 L stainless steel (SS) wires in terms of fixation rigidity, critical load, first/ultimate failure characteristics. The metal sutures were then immersed in Hank's balanced salt solution for 12 weeks immersion period, and their corrosion behaviours assessed.Zn–Cu wires showed similar fixation rigidity at 70.89 ± 6.97 N/mm as SS, but the critical load, first failure and ultimate failure characteristics were inferior to SS. The key challenges that limited the fixation effect of the Zn–Cu wires were poor mechanical strength, short elastic region, and strain softening behaviours, which resulted in poor load-bearing capabilities and reduced the knot security of the sutures. The in-situ biodegradation of the Zn–Cu suture was accompanied by the early onset of localised corrosion within the twisted knot and the section located next to the incision line. Crevice corrosion and strain-induced corrosion were the dominant mechanisms in the observed localised corrosion. The localised corrosion on the Zn–Cu sutures did not lead to a significant shift in fixation rigidity, critical load and the first failure characteristics. The findings suggest that the Zn-based biodegradable metallic wires could be a promising sternum closure suture material once the limitations in mechanical characteristics are addressed.

Oxygen transport through epoxy-based powder coatings in humid environments

Epoxies are used in various industrial applications as corrosion barriers but quantitative time-dependent predictions of their ability to mitigate corrosion attack at the metal/coating interface remain elusive. Permeability data for water and oxygen through epoxy-based coatings are of particular interest because the coating's barrier performance in humid environments is directly related to how quickly these reactants get to the metal/coating interface. There is, however, an absence of literature data to explain oxygen transport within coatings in the presence of condensed water and its associated plasticization effects. We show that water is a dominant player in the barrier performance of epoxy coatings because it blocks the transport of other permeants. Through analysis of the sorption isotherms and empirical permeation data, we determined adjustable parameters that explain water vapor plasticization effects in fusion bonded epoxy (FBE) at 65 °C. At this temperature, evolution of cavity formation and break-up of water clusters result in high mobility of water molecules inside the epoxy network. We also modeled oxygen transport through FBE in wet-state conditions based on time lag measurements, and with reported literature data on vapor and gas sorption in epoxy. In a mixed gas/vapor system, condensed water in FBE blocks and/or significantly decreases gas permeation in the coating. If the service temperature is low (<40 °C), water immobilizes the oxygen gas within microvoid regions in the glassy epoxy. Our experimental measurements, combined with Freeman's theoretical model for upper bound limits, showed that this water-induced blocking mechanism is sufficient to suppress corrosion reactions on the underlying substrate material. At 65 °C and above, the synergistic effect of coating plasticization by water molecules and dissolution of oxygen in a mobile water phase results in significant gaseous transport. We applied mathematical models based on proven sorption and transport models to FBE free-standing films and derived adjustable parameters to quantitatively explain this competitive permeation. Our O2 transport data show that, compared to the single-layer FBE, epoxy-based coatings with additional polyolefin layers can improve the barrier performance by deprivation of micropore channels for gas transport.

Erosion Attack: Harnessing Corruption To Improve Adversarial Examples.

Although adversarial examples pose a serious threat to deep neural networks, most transferable adversarial attacks are ineffective against black-box defense models. This may lead to the mistaken belief that adversarial examples are not truly threatening. In this paper, we propose a novel transferable attack that can defeat a wide range of black-box defenses and highlight their security limitations. We identify two intrinsic reasons why current attacks may fail, namely data-dependency and network-overfitting. They provide a different perspective on improving the transferability of attacks. To mitigate the data-dependency effect, we propose the Data Erosion method. It involves finding special augmentation data that behave similarly in both vanilla models and defenses, to help attackers fool robustified models with higher chances. In addition, we introduce the Network Erosion method to overcome the network-overfitting dilemma. The idea is conceptually simple: it extends a single surrogate model to an ensemble structure with high diversity, resulting in more transferable adversarial examples. Two proposed methods can be integrated to further enhance the transferability, referred to as Erosion Attack (EA). We evaluate the proposed EA under different defenses that empirical results demonstrate the superiority of EA over existing transferable attacks and reveal the underlying threat to current robust models. The source code is publicly available at https://github.com/mesunhlf/EA.